Merge: Update intel_idle to upstream 6.17

MR: https://gitlab.com/redhat/centos-stream/src/kernel/centos-stream-9/-/merge_requests/7315

JIRA: https://issues.redhat.com/browse/RHEL-113139

Some commits have been dropped; however, these have no functional
impact.

Signed-off-by: David Arcari <darcari@redhat.com>

Approved-by: Steve Best <sbest@redhat.com>
Approved-by: Tony Camuso <tcamuso@redhat.com>
Approved-by: CKI KWF Bot <cki-ci-bot+kwf-gitlab-com@redhat.com>

Merged-by: CKI GitLab Kmaint Pipeline Bot <26919896-cki-kmaint-pipeline-bot@users.noreply.gitlab.com>

arch/x86/kernel/smp.c

@@ -299,3 +299,27 @@ struct smp_ops smp_ops = {
 	.send_call_func_single_ipi = native_send_call_func_single_ipi,
 };
 EXPORT_SYMBOL_GPL(smp_ops);
+
+int arch_cpu_rescan_dead_smt_siblings(void)
+{
+	enum cpuhp_smt_control old = cpu_smt_control;
+	int ret;
+
+	/*
+	 * If SMT has been disabled and SMT siblings are in HLT, bring them back
+	 * online and offline them again so that they end up in MWAIT proper.
+	 *
+	 * Called with hotplug enabled.
+	 */
+	if (old != CPU_SMT_DISABLED && old != CPU_SMT_FORCE_DISABLED)
+		return 0;
+
+	ret = cpuhp_smt_enable();
+	if (ret)
+		return ret;
+
+	ret = cpuhp_smt_disable(old);
+
+	return ret;
+}
+EXPORT_SYMBOL_GPL(arch_cpu_rescan_dead_smt_siblings);

arch/x86/power/hibernate.c

@@ -188,7 +188,8 @@ out:
 
 int arch_resume_nosmt(void)
 {
-	int ret = 0;
+	int ret;
+
 	/*
 	 * We reached this while coming out of hibernation. This means
 	 * that SMT siblings are sleeping in hlt, as mwait is not safe
@@ -202,18 +203,10 @@ int arch_resume_nosmt(void)
 	 * Called with hotplug disabled.
 	 */
 	cpu_hotplug_enable();
-	if (cpu_smt_control == CPU_SMT_DISABLED ||
-			cpu_smt_control == CPU_SMT_FORCE_DISABLED) {
-		enum cpuhp_smt_control old = cpu_smt_control;
 
-		ret = cpuhp_smt_enable();
+	ret = arch_cpu_rescan_dead_smt_siblings();
-		if (ret)
+
-			goto out;
-		ret = cpuhp_smt_disable(old);
-		if (ret)
-			goto out;
-	}
-out:
 	cpu_hotplug_disable();
+
 	return ret;
 }

drivers/acpi/internal.h

@@ -174,6 +174,12 @@ bool processor_physically_present(acpi_handle handle);
 static inline void acpi_early_processor_control_setup(void) {}
 #endif
 
+#ifdef CONFIG_ACPI_PROCESSOR_CSTATE
+void acpi_idle_rescan_dead_smt_siblings(void);
+#else
+static inline void acpi_idle_rescan_dead_smt_siblings(void) {}
+#endif
+
 /* --------------------------------------------------------------------------
                                   Embedded Controller
    -------------------------------------------------------------------------- */

drivers/acpi/processor_driver.c

@@ -298,6 +298,9 @@ static int __init acpi_processor_driver_init(void)
 	 * after acpi_cppc_processor_probe() has been called for all online CPUs
 	 */
 	acpi_processor_init_invariance_cppc();
+
+	acpi_idle_rescan_dead_smt_siblings();
+
 	return 0;
 err:
 	driver_unregister(&acpi_processor_driver);

drivers/acpi/processor_idle.c

@@ -24,6 +24,8 @@
 #include <acpi/processor.h>
 #include <linux/context_tracking.h>
 
+#include "internal.h"
+
 /*
  * Include the apic definitions for x86 to have the APIC timer related defines
  * available also for UP (on SMP it gets magically included via linux/smp.h).
@@ -55,6 +57,12 @@ struct cpuidle_driver acpi_idle_driver = {
 };
 
 #ifdef CONFIG_ACPI_PROCESSOR_CSTATE
+void acpi_idle_rescan_dead_smt_siblings(void)
+{
+	if (cpuidle_get_driver() == &acpi_idle_driver)
+		arch_cpu_rescan_dead_smt_siblings();
+}
+
 static
 DEFINE_PER_CPU(struct acpi_processor_cx * [CPUIDLE_STATE_MAX], acpi_cstate);
 

drivers/idle/intel_idle.c

@@ -48,9 +48,11 @@
 #include <trace/events/power.h>
 #include <linux/sched.h>
 #include <linux/sched/smt.h>
+#include <linux/mutex.h>
 #include <linux/notifier.h>
 #include <linux/cpu.h>
 #include <linux/moduleparam.h>
+#include <linux/sysfs.h>
 #include <asm/cpu_device_id.h>
 #include <asm/intel-family.h>
 #include <asm/mwait.h>
@@ -91,9 +93,15 @@ struct idle_cpu {
 	 */
 	unsigned long auto_demotion_disable_flags;
 	bool disable_promotion_to_c1e;
+	bool c1_demotion_supported;
 	bool use_acpi;
 };
 
+static bool c1_demotion_supported;
+static DEFINE_MUTEX(c1_demotion_mutex);
+
+static struct device *sysfs_root __initdata;
+
 static const struct idle_cpu *icpu __initdata;
 static struct cpuidle_state *cpuidle_state_table __initdata;
 
@@ -142,8 +150,8 @@ static __always_inline int __intel_idle(struct cpuidle_device *dev,
 					int index, bool irqoff)
 {
 	struct cpuidle_state *state = &drv->states[index];
-	unsigned long eax = flg2MWAIT(state->flags);
+	unsigned int eax = flg2MWAIT(state->flags);
-	unsigned long ecx = 1*irqoff; /* break on interrupt flag */
+	unsigned int ecx = 1*irqoff; /* break on interrupt flag */
 
 	mwait_idle_with_hints(eax, ecx);
 
@@ -216,9 +224,9 @@ static __cpuidle int intel_idle_xstate(struct cpuidle_device *dev,
 static __cpuidle int intel_idle_s2idle(struct cpuidle_device *dev,
 				       struct cpuidle_driver *drv, int index)
 {
-	unsigned long ecx = 1; /* break on interrupt flag */
 	struct cpuidle_state *state = &drv->states[index];
-	unsigned long eax = flg2MWAIT(state->flags);
+	unsigned int eax = flg2MWAIT(state->flags);
+	unsigned int ecx = 1; /* break on interrupt flag */
 
 	if (state->flags & CPUIDLE_FLAG_INIT_XSTATE)
 		fpu_idle_fpregs();
@@ -1548,18 +1556,21 @@ static const struct idle_cpu idle_cpu_gmt __initconst = {
 static const struct idle_cpu idle_cpu_spr __initconst = {
 	.state_table = spr_cstates,
 	.disable_promotion_to_c1e = true,
+	.c1_demotion_supported = true,
 	.use_acpi = true,
 };
 
 static const struct idle_cpu idle_cpu_gnr __initconst = {
 	.state_table = gnr_cstates,
 	.disable_promotion_to_c1e = true,
+	.c1_demotion_supported = true,
 	.use_acpi = true,
 };
 
 static const struct idle_cpu idle_cpu_gnrd __initconst = {
 	.state_table = gnrd_cstates,
 	.disable_promotion_to_c1e = true,
+	.c1_demotion_supported = true,
 	.use_acpi = true,
 };
 
@@ -1598,12 +1609,14 @@ static const struct idle_cpu idle_cpu_snr __initconst = {
 static const struct idle_cpu idle_cpu_grr __initconst = {
 	.state_table = grr_cstates,
 	.disable_promotion_to_c1e = true,
+	.c1_demotion_supported = true,
 	.use_acpi = true,
 };
 
 static const struct idle_cpu idle_cpu_srf __initconst = {
 	.state_table = srf_cstates,
 	.disable_promotion_to_c1e = true,
+	.c1_demotion_supported = true,
 	.use_acpi = true,
 };
 
@@ -1664,7 +1677,7 @@ static const struct x86_cpu_id intel_idle_ids[] __initconst = {
 };
 
 static const struct x86_cpu_id intel_mwait_ids[] __initconst = {
-	X86_MATCH_VENDOR_FAM_FEATURE(INTEL, 6, X86_FEATURE_MWAIT, NULL),
+	X86_MATCH_VENDOR_FAM_FEATURE(INTEL, X86_FAMILY_ANY, X86_FEATURE_MWAIT, NULL),
 	{}
 };
 
@@ -2323,6 +2336,88 @@ static void __init intel_idle_cpuidle_devices_uninit(void)
 		cpuidle_unregister_device(per_cpu_ptr(intel_idle_cpuidle_devices, i));
 }
 
+static void intel_c1_demotion_toggle(void *enable)
+{
+	unsigned long long msr_val;
+
+	rdmsrl(MSR_PKG_CST_CONFIG_CONTROL, msr_val);
+	/*
+	 * Enable/disable C1 undemotion along with C1 demotion, as this is the
+	 * most sensible configuration in general.
+	 */
+	if (enable)
+		msr_val |= NHM_C1_AUTO_DEMOTE | SNB_C1_AUTO_UNDEMOTE;
+	else
+		msr_val &= ~(NHM_C1_AUTO_DEMOTE | SNB_C1_AUTO_UNDEMOTE);
+	wrmsrl(MSR_PKG_CST_CONFIG_CONTROL, msr_val);
+}
+
+static ssize_t intel_c1_demotion_store(struct device *dev,
+				       struct device_attribute *attr,
+				       const char *buf, size_t count)
+{
+	bool enable;
+	int err;
+
+	err = kstrtobool(buf, &enable);
+	if (err)
+		return err;
+
+	mutex_lock(&c1_demotion_mutex);
+	/* Enable/disable C1 demotion on all CPUs */
+	on_each_cpu(intel_c1_demotion_toggle, (void *)enable, 1);
+	mutex_unlock(&c1_demotion_mutex);
+
+	return count;
+}
+
+static ssize_t intel_c1_demotion_show(struct device *dev,
+				      struct device_attribute *attr, char *buf)
+{
+	unsigned long long msr_val;
+
+	/*
+	 * Read the MSR value for a CPU and assume it is the same for all CPUs. Any other
+	 * configuration would be a BIOS bug.
+	 */
+	rdmsrl(MSR_PKG_CST_CONFIG_CONTROL, msr_val);
+	return sysfs_emit(buf, "%d\n", !!(msr_val & NHM_C1_AUTO_DEMOTE));
+}
+static DEVICE_ATTR_RW(intel_c1_demotion);
+
+static int __init intel_idle_sysfs_init(void)
+{
+	int err;
+
+	if (!c1_demotion_supported)
+		return 0;
+
+	sysfs_root = bus_get_dev_root(&cpu_subsys);
+	if (!sysfs_root)
+		return 0;
+
+	err = sysfs_add_file_to_group(&sysfs_root->kobj,
+				      &dev_attr_intel_c1_demotion.attr,
+				      "cpuidle");
+	if (err) {
+		put_device(sysfs_root);
+		return err;
+	}
+
+	return 0;
+}
+
+static void __init intel_idle_sysfs_uninit(void)
+{
+	if (!sysfs_root)
+		return;
+
+	sysfs_remove_file_from_group(&sysfs_root->kobj,
+				     &dev_attr_intel_c1_demotion.attr,
+				     "cpuidle");
+	put_device(sysfs_root);
+}
+
 static int __init intel_idle_init(void)
 {
 	const struct x86_cpu_id *id;
@@ -2373,6 +2468,8 @@ static int __init intel_idle_init(void)
 		auto_demotion_disable_flags = icpu->auto_demotion_disable_flags;
 		if (icpu->disable_promotion_to_c1e)
 			c1e_promotion = C1E_PROMOTION_DISABLE;
+		if (icpu->c1_demotion_supported)
+			c1_demotion_supported = true;
 		if (icpu->use_acpi || force_use_acpi)
 			intel_idle_acpi_cst_extract();
 	} else if (!intel_idle_acpi_cst_extract()) {
@@ -2386,6 +2483,10 @@ static int __init intel_idle_init(void)
 	if (!intel_idle_cpuidle_devices)
 		return -ENOMEM;
 
+	retval = intel_idle_sysfs_init();
+	if (retval)
+		pr_warn("failed to initialized sysfs");
+
 	intel_idle_cpuidle_driver_init(&intel_idle_driver);
 
 	retval = cpuidle_register_driver(&intel_idle_driver);
@@ -2404,17 +2505,20 @@ static int __init intel_idle_init(void)
 	pr_debug("Local APIC timer is reliable in %s\n",
 		 boot_cpu_has(X86_FEATURE_ARAT) ? "all C-states" : "C1");
 
+	arch_cpu_rescan_dead_smt_siblings();
+
 	return 0;
 
 hp_setup_fail:
 	intel_idle_cpuidle_devices_uninit();
 	cpuidle_unregister_driver(&intel_idle_driver);
 init_driver_fail:
+	intel_idle_sysfs_uninit();
 	free_percpu(intel_idle_cpuidle_devices);
 	return retval;
 
 }
-device_initcall(intel_idle_init);
+subsys_initcall_sync(intel_idle_init);
 
 /*
  * We are not really modular, but we used to support that.  Meaning we also

include/linux/cpu.h

@@ -116,6 +116,7 @@ extern void cpu_maps_update_begin(void);
 extern void cpu_maps_update_done(void);
 int bringup_hibernate_cpu(unsigned int sleep_cpu);
 void bringup_nonboot_cpus(unsigned int setup_max_cpus);
+int arch_cpu_rescan_dead_smt_siblings(void);
 
 #else	/* CONFIG_SMP */
 #define cpuhp_tasks_frozen	0
@@ -130,6 +131,8 @@ static inline void cpu_maps_update_done(void)
 
 static inline int add_cpu(unsigned int cpu) { return 0;}
 
+static inline int arch_cpu_rescan_dead_smt_siblings(void) { return 0; }
+
 #endif /* CONFIG_SMP */
 extern struct bus_type cpu_subsys;